Current Events > Analysis of full range of GRE scores shows no evidence of prediction of outcomes

(+) Yeah! (+)

https://www.biorxiv.org/content/early/2018/07/20/373225

Not too surprising. I don't think many people would honestly predict an SAT 2.0 to predict success in PhD programs. Yet, we still cling to it like an irrational tradition.
---

Most standardized tests are garbage. They tend to predict how well you'd do at other standardized tests than how successful you'll be in a selected field.
---

Yeah, I agree. I mostly think they exist as another place where relative privilege can assert it's little nose.

Unfortunately, we really like metrics for decisions like admittance and whatnot. Even if they are flawed. They give the process a feeling of fake certainty and objectitivity.
---

Are there no practical application test?
---

its kinda like when congress was talking to that dude about pulling people's credit scores when they hire them.

congress was like "is there proof that bad credit scores make bad employees?"
and he was like "no"
and they said "then why do you do let people do it"
and he was like "so you can use that information to make your own decision"

HUHHH? so theres no proof it means anything but you're free to pretend there is.

fuck this country.
---

E32005 posted...

Are there no practical application test?

What do you mean?
---

COVxy posted...

E32005 posted...

Are there no practical application test?

What do you mean?

Hes saying you have to test the applicants some way. And testing them on material they were supposed to have learned is fair enough to get a decent overview.
---

Russman posted...

COVxy posted...

E32005 posted...

Are there no practical application test?

What do you mean?

Hes saying you have to test the applicants some way. And testing them on material they were supposed to have learned is fair enough to get a decent overview.

I mean, I got the general idea of what he meant, but I wanted to know more specifically what he meant.

Cuz there are subject GREs, but they are stupid as well. The closest to an honest "practical" is the interview, where the PI gets to suss out how much of the student's CV is of their own making. But interviews are late stage, it is often the case that if you got the interview you are already essentially accepted.
---

It's actually important to note that the low sample size here means that the actual correlations are estimated at an extremely low level of precision, however, large range of GRE scores means that if the actual population correlation value did fall within the scope of reasonable estimates, it would still be an unreasonably poor predictor.
---

what do you think about eric and sonia vallabh's quest to cure prion diseases
---